Summary: Averaging brain-scan data across large groups can mislead researchers about how individual human brains function. By examining functional MRI data from more than 4,000 children at the single-subject level, Stanford researchers found that many children — especially those who struggle with inhibitory control — show brain dynamics that often run opposite to group-average patterns.
This work challenges longstanding assumptions in cognitive neuroscience and supports a shift toward personalized approaches in psychiatry and treatments for conditions such as ADHD.
Key Facts
- The speed–accuracy paradox: Group-level relationships can mask individual behavior. While faster individuals in a group may also be more accurate, a single person typically slows down to improve accuracy. Brain patterns likewise differ when analyzed across people versus within one person over time.
- Default Mode Network (DMN) reversal: At the group level, slower reaction times were associated with increased DMN activity (linked to mind-wandering). Within individuals, slower reactions commonly corresponded to decreased DMN activity — a reversal of the group pattern.
- Proactive and reactive control are distinct: Cognitive control is not a single monolithic ability but a set of coordinated sub-processes. Proactive control (preparing to stop) and reactive control (executing the stop) are represented by different neural mechanisms.
- Multiple neural pathways to success: Children with weaker cognitive control often rely on alternative neural strategies, suggesting that inhibitory control is a flexible skill that can be regulated or learned through different routes rather than a fixed capacity.
Source: Stanford
Studying cognition by averaging brain scans hides how individuals actually use their brains, a new Stanford Medicine study shows.
When researchers analyzed each child’s functional MRI (fMRI) data individually, distinct patterns emerged that were obscured by conventional group-averaging methods. These individual dynamics are particularly informative for children who struggle with goal-directed behavior and inhibitory control.
The findings, relevant to understanding the neural basis of disorders such as attention-deficit/hyperactivity disorder (ADHD), were published in Nature Communications on April 27.
“Examining how neural dynamics unfold within individual brains gives important insights into individual differences that group approaches cannot reveal,” said Percy Mistry, PhD, research scholar in psychiatry and behavioral sciences and a lead author of the study.
Lead authorship is shared by Percy Mistry and Nicholas Branigan, MS, with Vinod Menon, PhD, serving as senior author. The research focused on inhibitory cognitive control — the brain’s ability to suppress distractions and irrelevant impulses so a person can pursue a goal — in more than 4,000 children.
Researchers compared standard group-average analyses with trial-by-trial, within-subject analyses that captured how each child’s brain activity changed while repeating the same task. This single-subject perspective allowed the team to identify subgroups of children who differed in cognitive control and performance monitoring — the ability to adjust strategy after an error.
For example, children with strong cognitive control and performance monitoring exhibited brain dynamics that were often completely different from those with weaker regulation. “Group averages can fundamentally mischaracterize how the brain dynamically governs behavior,” Menon said.
A clue from the speed–accuracy trade-off
Psychologists have long known that correlations observed across people do not always reflect relations within a single person. The prototypical example is the speed–accuracy trade-off: across a group, faster responders may appear more accurate, but within one person, a push for speed typically reduces accuracy.
The Stanford team tested whether a similar divergence exists between group-level and individual-level brain–behavior links. Using fMRI data collected while children performed a stop-signal task (a standard measure of inhibitory control), the researchers compared across-subject trends with within-subject trial dynamics.
In the stop-signal task, children press a button when they see “Go” but must withhold the response if an infrequent, unpredictable “Stop” cue appears. The team examined brain activity on every trial and contrasted between-subject and within-subject associations.
Group-level analyses showed that slower reaction times correlated with increased activity in several regions, including the default mode network. In contrast, within-subject analyses revealed that, for an individual, a slower reaction was typically accompanied by decreased DMN activity — a direct reversal of the group pattern.
“Group-level associations can substantially misrepresent the neural dynamics that underlie an individual’s processing speed,” the authors write.
Using a Bayesian computational model of cognitive dynamics, the team also studied how children adjusted their responses across repeated trials. Some children adaptively increased their readiness to stop after an initial “Stop,” showing faster stopping reactions on subsequent trials. Others displayed maladaptive adjustments, anticipating fewer stops and showing the opposite behavioral and neural pattern.
The authors note that several effects seen in the full sample were driven primarily by one subgroup, so averaging obscured what was true for many individuals.
Multiple pathways and implications for treatment
Analyses showed that proactive and reactive control are supported by different brain regions and that their interactions vary across children. When a child has weaker control in one component, they may compensate by strengthening another component or recruiting an alternate neural route.
“This reframes cognitive control as a flexible set of strategies rather than a single static capacity,” Mistry said. That perspective could guide classroom and clinical strategies for children with ADHD, tailoring interventions to their specific neural strengths and weaknesses — for example, relying on proactive preparation when reactive stopping is weaker.
Menon added that neuroscience should pay closer attention to each person’s unique neural responses, especially when the goal is to understand or modify behavior in real time. “There is no single average brain. We need to ask: how is this child or adult responding to particular situations that demand attention and adaptive regulation?”
The study used baseline data from the Adolescent Brain and Cognitive Development (ABCD) study and underscores the importance of distinguishing between-subjects and within-subjects inferences in neuroscience. These results have implications for understanding cognitive mechanisms and designing personalized interventions.
Data and funding
The ABCD study data used in this research are archived at the National Institute of Mental Health Data Archive. The ABCD study is supported by the National Institutes of Health and federal partners under multiple awards. At Stanford Medicine, this work was supported by NIH grants MH121069 and MH124816, NSF grant 2024856, and the Stanford Maternal and Child Health Research Institute; Stanford University and Stanford Research Computing provided computational resources.
Key Questions Answered:
A: Averaging produces a composite “average brain” that does not exist in any individual. That average can mask important differences in how children — including those with ADHD or mood disorders — process information and regulate behavior.
A: Rather than labeling a child as having uniformly “poor focus,” clinicians could identify which control pathway is strongest or weakest. Interventions and classroom strategies could then support a child’s strengths (for example, proactive preparation) while addressing specific weaknesses.
A: Not wrong, but incomplete. Group studies reveal important general trends, but they can fail to capture or even reverse the patterns that appear within an individual. This study clarifies when and why between-subject and within-subject inferences diverge.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The journal paper was reviewed in full.
- Additional context was added by staff.
About this neuroscience and ADHD research news
Author: Erin Digitale
Source: Stanford
Contact: Erin Digitale, Stanford
Image: Image credit: Neuroscience News
Original Research: “Nonergodicity and Simpson’s paradox in neurocognitive dynamics of cognitive control” by Percy K. Mistry, Nicholas K. Branigan, Zhiyao Gao, Weidong Cai & Vinod Menon. Nature Communications. DOI: 10.1038/s41467-026-71404-0. The paper is open access.
Abstract
Nonergodicity and Simpson’s paradox in neurocognitive dynamics of cognitive control
Nonergodicity and Simpson’s paradox present significant but underappreciated challenges in cognitive neuroscience. Using brain imaging and behavioral data from over 4,000 individuals combined with a Bayesian model of cognitive dynamics, the study examined brain–behavior relationships at both between-subjects and within-subjects levels.
The results revealed pervasive nonergodicity: brain–behavior associations often reversed across levels of analysis. Within-subject analyses uncovered separate neural representations for reactive and proactive control and showed that individuals who adaptively versus maladaptively regulated cognitive control exhibited distinct brain–behavior associations.
These findings demonstrate that between-subject analyses can fundamentally mischaracterize within-individual mechanisms. Group-level patterns not only disagreed with individual-level patterns but frequently reversed them. The study highlights the importance of distinguishing between-subject and within-subject inferences in neuroscience and the implications for understanding cognitive mechanisms and designing personalized interventions.